RICE PLANT DISEASE IDENTIFICATION AND DETECTION TECHNOLOGY THROUGH CLASSIFICATION OF MICROORGANISMS USING FUZZY NEURAL NETWORK

Abstract

This paper describes a method of using sound signal processing system to efficiently detect and identify the three common microorganisms that cause diseases in the rice farmland of the Philippines: (1) Xanthomonas oryzae, (2) Thanatephorus cucumeris and (3) Magnaporthe oryzae. Sound signals from samples of rice leaves infected by the above mentioned bacteria were recorded using a designed anechoic chamber through an electret condenser microphone and were processed via spectral subtraction to eliminate the effects of noise. Mel Frequency Cepstral Coefficient was used to extract the needed features of each input for the ANFIS learning algorithm. The Fuzzy neural network was applied to train the system based on 450 recorded sound data where 80% were used for training and 20% for testing. A program was also developed that will generate a report in PDF format showing the diagnosis and curing methods for the infected sample to prevent its further infestation. Test results showed recognition accuracy of the bacteria, Xanthomonas oryzae, Magnaporthe oryzae, and Thanatephorus cucumeris, of 93.33%, 100% and 96.67% repectively.